Boost C++ Libraries of the most highly regarded and expertly designed C++ library projects in the world. Herb Sutter and Andrei Alexandrescu, C++ Coding Standards


Whittaker-Shannon interpolation


#include <boost/math/interpolators/whittaker_shannon.hpp>
namespace boost { namespace math { namespace interpolators {

  template <class RandomAccessContainer>
  class whittaker_shannon

      using Real = RandomAccessContainer::value_type;

      whittaker_shannon(RandomAccessContainer&& v, Real left_endpoint, Real step_size);

      Real operator()(Real x) const;

      Real prime(Real x) const;

}}} // namespaces

Whittaker-Shannon Interpolation

The Whittaker-Shannon interpolator takes equispaced data and interpolates between them via a sum of sinc functions. This interpolation is stable and infinitely smooth, but has linear complexity in the data, making it slow relative to compactly-supported b-splines. In addition, we cannot pass an infinite amount of data into the class, and must truncate the (perhaps) infinite sinc series to a finite number of terms. Since the sinc function has slow 1/x decay, the truncation of the series can incur large error. Hence this interpolator works best when operating on samples of compactly-supported functions. Here is an example of interpolating a smooth "bump function":

auto bump = [](double x) { if (std::abs(x) >= 1) { return 0.0; } return std::exp(-1.0/(1.0-x*x)); };

double t0 = -1;
size_t n = 2049;
double h = 2.0/(n-1.0);

std::vector<double> v(n);
for(size_t i = 0; i < n; ++i) {
    double t = t0 + i*h;
    v[i] = bump(t);

auto ws = whittaker_shannon(std::move(v), t0, h);

double y = ws(0.3);

The derivative of the interpolant can also be evaluated, but the accuracy is not as high:

double yp =;

Complexity and Performance

The call to the constructor requires 𝑶(1) operations, simply moving data into the class. Each call to the interpolant is 𝑶(n), where n is the number of points to interpolate.